URONIC ACID
IN
HEPARINS AND
HEPARAN
SULFATES
Uronic Acid Composition of Heparins and Heparan Sulfatest R . L. Taylor,$ J. E. Shively, H. E. Conrad,* and J. A. Cifonelli
ABSTRACT: The L-iduronic and D-glucuronic acid contents of a variety of heparin and heparan sulfate samples have been determined by measuring the amounts of 3H recovered in glucose, idose, and idosan when polymers are carboxyl reduced with sodium [ 3H]borohydride. The percentage of the total uronic acid represented by L-iduronic acid varied from -50 to 90 in heparins and from 30 to 55 in heparan sul-
F
ollowing the initial demonstration of the presence of Dglucuronic acid in heparin (Wolfrom and Rice, 1946) and heparan sulfate (Brown, 1957), Brown et a/. (1961) presented evidence that heparin contained, in addition to D-glucuronic acid, another type of uronic acid, thought to be a ketouronic acid. This heparin component wa's identified as L-iduronic acid by Cifonelli and Dorfman (1962) who showed that it was present in heparan sulfate and mactins as well. Wolfrom et al. (1969a) have isolated L-iduronic acid from heparin as its crystalline brucinium salt, and Perlin et al. (1970) have established by nuclear magnetic resonance (nmr) techniques that Liduronic acid is a major fraction of the total uronic acid in heparin. Heparins, which show anticoagulant or lipoprotein lipase activation activities, are highly N- and 0-sulfated, while heparan sulfates, which are less highly sulfated, are diminished in these activities (Grossman et al., 1971). Periodate oxidation studies have shown that the L-iduronic acid, but not the D-glucuronic acid, residues of heparin are 0sulfated at C-2 (Wolfrom et al., 1969b; Lindahl and Axelsson, 1971). Consequently, the L-iduronic acid residues carry a significant fraction of the total 0-sulfate in heparin, The possibility exists, therefore, that the decreased sulfation (and biological activity) in heparan sulfates might be accompanied by a corresponding decrease in L-iduronic acid contents. This paper reports the analyses of the L-iduronic acid contents of a series of previously described heparin and heparan sulfate preparations and correlates these data with those for other variable constituents of this group of polysaccharides. Experimental Procedures Methods. Beef lung heparin and heparin by-products were obtained from Dr. L. L. Coleman, The Upjohn Co. Similar preparations from hog mucosa and heparin from beef mucosa were provided by Dr. H. H. R. Weber, The Wilson Labora-
t From the Department of Biochemistry, University of Illinois, Urbana, Illinois 61801 ( R . L. T.,J. E. S., and H. E. C.), and the Departments of Pediatrics and Biochemistry, The LaRabida-University of Chicago Institute, and The Joseph P. Kennedy, Jr., Mental Retardation Center, University of Chicago, Chicago, Illinois 60637 (J. A. C.). Supported by grants from the U. S. Public Health Service, National Institute of Allergy and Infectious Diseases (AI 09551), National Institute for Arthritis and Metabolic Diseases (AM 05996), and National Heart Institute (HE-12927). $ Present address: University of Chicago Medical School, Chicago, Ill. 60615.
fates. Ratios of N- and 0-sulfate in these polymers generally increased with increasing L-iduronic acid content. The Nsulfate :D-glucosamineratios varied from 0.7 to 1.O for heparins and from 0.3 to 0.6 for heparan sulfates. The 0 - s u l f a t e n glucosamine ratios ranged from 0.9 to 1.5 for heparins and 0.2 to 0.8 for heparan sulfates.
tories. Purified heparin fractions described previously (Cifonelli and King, 1970a) include those from beef lung (BLH I), bovine mucosa (BHM I), and whale tissue (WH I). Hog mucosa heparins include those obtained from Sigma (HMH I) and from heparin or heparin by-products after fractionation with cetylpyridinium chloride or on Dowex 1 (chloride) columns (HMH 11, BLH 11, and BLH 111, RodCn et al., 1972). Mactin (M I) was isolated from clam tissues and purified with cetypyridinium chloride (Cifonelli and Mathews, 1972a). Heparan sulfate fractions were isolated from beef lung byproducts (HSB 1-111) and from hog mucosa by-products (HSH 1-111) by fractionation on Dowex 1 columns (Rodin et al., 1972). Umbilical cord heparan sulfate (HSU I) was reported earlier (Cifonelli and King, 1970b). The heparan sulfate degradation fraction (HSD I) was produced by degradation of hog mucosa heparan sulfate (HSH I) with nitrites and isolation of nonreacting N-acetylated glucosamine sections of estimated molecular size 3-4 x lo3 by gel filtration on Sephadex G-25 as described previously (Cifonelli, 1968). A disulfated disaccharide fraction (D I), composed of uronic acid and 2,5-anhydromannose units, was obtained after reaction of hog mucosa heparin with alkyl nitrites and fractionation of the products on Dowex 1 and Sephadex G-25 as described previously (Cifonelli and King, 1972). Analytical procedures for estimating uronic acid, hexosamine, total sulfate, N-sulfate, neutral sugars, and amino acids have been described previously (Lindahl et al., 1965). Anticoagulant assays were performed by Dr. L. W. Van Ness, the Wilson Laboratories. Molecular weights were approximated by measuring elution volumes obtained by gel filtration of samples on a Sephadex G-75 column (0.85 x 185 cm) and comparing the values with a molecular weight diagram given by heparin and heparan sulfate fractions of known sizes, ranging from -4.5 to 11 x lo3, as described by Constantopoulos et al. (1969) and Wasteson (1969). The molecular weights of HSU I and M I were estimated from 7 measurements. Quantitation of D - G l u c u r o n i c and L-Iduronic Acid Contents. The reaction sequence used for the quantitation of uronic acids is based upon a recently described procedure for quantitative carboxyl reduction and stoichiometric depolymerization of glycosaminoglycuronans (Taylor and Conrad, 1972) and is illustrated in Figure 1. In this procedure the polymers (I) are allowed to react with l-ethyl-3-(dimethylaminopropyl)carbodiimide and the carboxyl-activated polymers are reduced with sodium [ 3H]borohydride (reaction 1). The reduction B I O C H E M I S T R Y , VOL.
12,
NO.
19, 1 9 7 3
3633
T A Y L O R e t al.
product, which is labeled with two tritiums per reduced carboxyl group, is then stoichiometrically depolymerized by acid hydrolysis (reaction 2) followed by deaminative cleavage of the remaining glucosaminides with nitrous acid (reaction 3) as described earlier. The monosaccharides thus formed are reduced with unlabeled sodium borohydride (reaction 4) and quantitated by radiochromatography (Conrad et id., 1973). This reaction sequence yields a mixture of L-[3H]idosan (11) and L[ 3H]iditol (111) from L-iduronic acid (Perlin and Sanderson, 1970), D-[ 3H]glucitol (V) from D-glucuronic acid, and unlabeled anhydromannitol (IV). The total 3H counts appearing in these three peaks is equivalent to the total uronic acid; the 3H counts in the L-iditol and L-idosan peaks are summed as a measure of the total equivalents of L-iduronic acid. Quantitative recoveries of the monosaccharides from the starting polysaccharide are obtained on the radiochromatograms (Taylor and Conrad, 1972). A typical reaction sequence is carried out as follows. A mucopolysaccharide sample (7 mg) is allowed to react a t room temperature with 19.2 mg of 1-ethyl-3-(dimethylaminopropyl)carbodiimide in 1 ml of water a t pH 4.75 for 1 hr. An aliquot (125 pi) of the reaction mixture is added to 250 pi of 3 hi sodium [ 3H]borohydride (14.3 mCi,'mmol, New England Nuclear Corporation) and the mixture is heated at 50" for 2 hr). The sample is then cooled and excess borohydride is destroyed by acidification with 3 N sulfuric acid. Salts are removed by dialysis and the dialyzed sample is transferred t o a test tube and evaporated to dryness in a stream of air at 50". The [3HH]carboxyl-reducedproduct is taken up in 50 p1 of water, 25 pl each of [lE]glucose solution (20,000 cpm/pl, 200 mCilmmol) and 4 N sulfuric acid is added, and the sample is hydrolyzed at 100" for 6 hr. An aliquot of the hydrolysate is deaminated by addition of sodium nitrite, aldehyde reduced with unlabeled sodium borohydride, and analyzed by radiochromatography (Shively and Conrad, 1970). Results and Discussion Table I presents the analytical data for those constituents of heparin and heparan sulfate which are relatively invariant for all fractions. The compositions of these preparations are in agreement with those generally reported for heparin and heparan sulfate fractions. For all samples the ratios of uronic acid to hexosamine are very similar regardless of source. The colorimetric methods used to obtain the percentages of uronic and hexosamine yield molar ratios of uronic acid to hexosamine considerably in excess of 1.0 in spite of a large accumulation of data which has led to the belief that the basic heparin structure is best represented by the repeating disaccharide, hexuronosylglucosamine (Jeanloz, 1970). Such a structure should give a hexuronic acid: glucosamine ratio of 1.0, after correction for the D-glucuronic acid residue of the linkage region (see below). As shown in Table I, ratios of hexuronic acid: hexosamine ranging from 1.2 to 1.5 are obtained in the radiochromatographic analyses of this series of heparins and heparan sulfates. Values of hexuronic acid in excess of hexosamine have also been reported by Perlin et a/. (1970) on the basis of the proton magnetic resonance spectrum of heparin. It may be noted that our previous application of these radiochromatographic procedures in the analysis of mLicopolysaccharides (Taylor and Conrad, 1972) yielded the expected hexuronic acid: hexosamine ratios of 1.0 for both hyaluronic acid and chondroitin sulfate. In these earlier studies it was also shown that quantitative recoveries of monosaccharides were obtained from a measured weight of muco-
3634
B I O C H E M I S T R Y , VOL.
12,
NO.
19, 1 9 7 3
OH
t HONO CHzOH
(31 C3H20H
0-
t H+
(2)
0N
0
N
$0;
40;
50;
so3
(1)
0-
1: The reaction sequence used in uronic acid analyses of heparins and heparan sulfates.
FIGURE
polysaccharide in the sequence of reactions shown in Figure 1. The very low levels of D-galactosamine in these preparations indicate lack of significant contamination with dermatan sulfate, commonly found in varying concentrations in less purified preparations. In the mactin fraction, however, approximately 15% of the hexosamine was found to be Dgalactosamine. This type of preparation was found to have a backbone similar to beef lung heparin, with 9 8 ~ ~ 9 9of % its glucosamine residues substituted with N-sulfate groups (Cifonelli and Mathews, 1972b). Since most of the galactosamine appears to be derived from spisulan (Cifonelli and Mathews, 1972b), in which the amino sugar is not associated with uronic acid, estimation of the percentage of each uronic acid in mactin is not appreciably affected by this contaminant. The presence of D-galactose and D-XylOSe in all of these samples reflects the presence of the linkage region. The linkage region also includes a D-glucuronic acid moiety (Lindahl and RodCn, 1965). This residue is determined as part of the total D-glucuronic acid in analysis for the relative amounts of Dglucuronic and L-iduronic acids and must be taken into consideration when it is wished to estimate the L-iduronic and Dglucuronic acid contents solely of the heparin backbone. Estimations of the D-glucuronic acid content originating from the linkage region were made on the basis of molecular weights, which ranged from 5.0 to 27 X lo3. These data are presented in Table 11. The values listed for linkage region Dglucuronic acid are maximal and assume that all molecules possess D-glucuronic acid at the linkage position. This is a reasonable assumption since most of the preparations contain appreciable amounts of D-galactose, which can only be derived from the linkage region, even in samples in which D-
URONIC
TABLE I :
ACID
IN HEPARINS AND HEPARAN
SULFATES
Composition of Heparin and Heparan Sulfate Preparations.
Fraction
Source
BLH I BLH I1 BLH I11 HMH I HMH I1 BMH I WH I MI HSH I HSH I1 HSH Ill HSB I HSB I1 HSB I l l HSU I HSD I D I
Beef lung Beef lungBeef lung Hog mucosa Hog mucosa Beef mucosa Whale Clam Hog mucosa by-product Hog mucosa by-product Hog mucosa by-product Beef lung by-product Beef lung by-product Beef lung by-product Human umbilical cord Heparan sulfate by-product Heparin disaccharide
Uronic Hexos- HexAb Acid" amine" ___ (%) (%) HexN 39 38 40
23 21 22
36 49 43 39 38 45 34 38 38 44 45 39 13
21 24 23 26 21 30 26 21 25 25 29 26 30'
1.4 1.5 1.3 1.4 1.5 1.4 1.4 1.2 1.3 1.3 1.3 1.2 1.2
Mol/mol of Glc-N Gal
Xyl
Gal N
Ser
GlY
0,007 0.005 0.030 0.031 0.080 0.050 0.011 0.080 0.042 0.011 0,010
0.003 0,001 0.021 0.020 0,042 0.035 0,004 0.020 0.022 0.006 0.005
Trace 0,004 Trace 0,010 0.022 0.025 Trace 0.150 0.016 0.012 0.008
Trace 0,013 Trace 0.006 0.005 0,044 Trace 0.028 0.017 0.018 0.006
Trace 0.009 Trace Trace 0.003 0.024 Trace 0.011 0.011 0.017 Trace
0.042 0,028 0.042
0.024 0.018 0.020
0.010 0.002 0.018
Trace Trace 0.019
Trace Trace 0.009
Molar ratio of total hexuronic acid to hexosamine determined by radioa Determined as described by Lindahl et al. (1965). chromatographic analysis (Conrad et ai.,1973). Present as 2,5-anhydromannose.
xylose and serine are present in low or negligible concentrations. The percentage of linkage region D-glucuronic acid is subtracted from the percentage of total uronic acid represented by D-glucuronic acid to obtain the value for the per cent Dglucuronic acid in the backbone of these polymers. The relative amounts of D-glucuronic and L-iduronic acids in the backbone may then be compared with the other variable parameters of this series of polysaccharides as shown in Table 11. For convenience the undegraded samples are grouped into heparin and heparan sulfate fractions, based primarily on the marked difference in their anticoagulant properties. When both groups are considered together, it is seen that the percentage of L-iduronic acid varies from 27 to 84, with values in the higher part of this range characterizing the heparins.' Similarly, the 0-sulfate values increase with increasing Nsulfate, and both N - and O-sulfate increase with increasing Liduronic acid. Curiously, the sharp delineation between heparins and heparan sulfates reflected in the anticoagulant activities is not so sharply defined in terms of sulfate content or per cent L-iduronic acid. Several fractions merit special comment. The mactin fraction [MI) has a relatively low L-iduronic acid content in spite of its high anticoagulant activity. This low level of L-iduronic acid was suggested previously (Cifonelli and Mathews, 1972a) on the basis of paper chromatographic evidence and ratios of uronic acid to 2,5-anhydromannose obtained from degradation products formed upon reaction with mactin with nitrous acid. Several degraded polymer fractions have been examined for their uronic acid content. Fraction BLH 111 was obtained as a minor fraction after chromatography of heparin on Dowex 1, chloride. This fraction was of unusually low mol Dr. Ulf Lindahl (personal communication) has determined the relative amounts of D-glucuronic acid and L-iduronic acid in a purified heparin from hog mucosa and a heparan sulfate from human aorta and has obtained results comparable to those presented in TableII.
lecular size, corresponding to a molecular weight of -5.2 X 102 daltons, and is estimated to have D-glucuronic acid in the linkage region amounting to 10% of the total uronic acid. Since this product possesses both N- and 0-sulfate in the range found for heparins, it appears that the fraction represents a polysaccharide derived from heparin. The presence of appreciable D-xylose in the sample indicates that the linkage region is not degraded to any major extent. Furthermore, the proportion of L-iduronic acid in this material is 75% of the total uronic acid, a value at the higher end for heparins, adding support for viewing this product as a heparin type. If all of the L-iduronic acid is 0-sulfated at C-2 as found for heparin (Wolfrom et ai.,1969b; Lindahl and Axelsson, 1971 ; R. L. Taylor and H. E. Conrad, unpublished results) and for at least some samples of heparan sulfate (J. E. Shively and H. E. Conrad, unpublished results), the data presented here show that in none of the samples is there sufficient 0-sulfate for all of the glucosamine residues to be 0-sulfated at C-6. That 0sulfate may vary on the N-acetyl-D-glucosamine-containing sections of heparan sulfates has been reported previously for heparan sulfates from umbilical cords (Cifonelli and King, 1970b) and beef lung (Cifonelli, 1968) and from rat brain (Margolis and Atherton, 1973). The present data suggest that the variability of the 0-sulfation of glucosamine residues is also observed in the heparin fractions. The two degraded fractions, HS DI and DI, recovered after nitrite treatment of heparan sulfate and heparin, respectively, show a further correlation of sulfate and L-iduronic acid contents. HS DI is a segment of heparan sulfate in which the glucosamine residues are N-acetylated. The low level of sulfate in this fraction suggested that D-glucuronic acid should predominate in this preparation; this is confirmed by the results in Table 11. Fraction DI, a di-0-sulfated disaccharide obtained by nitrite cleavage at the N-sulfated glucosamine residues of heparin (Cifonelli and King, 1972), consists of uronic acid and 2,5-anhydromannose only. Evidence from several sources indicates that both of these residues may be sulfated BIOCHEMISTRY V O, L . 1 2 , N O . 1 9 , 1 9 7 3
3635
TAYLOR
TABLE 11:
e t a[.
Variation of Iduronic Acid Content with Other Variable Physical, Structural, and Biological Parameters.
% of Total Uronic Acid Glc A Fraction Heparins BLH I BLH I1 BLH I11 HMH I H M H I1 BMH I WH I MI Heparan sulfates HSH I HSH I1 HSH 111 HSB I HSB I1 HSB I11 HSU I Degradation products HSD I DI
[alga
Mol Wt
(deg)
(X
43 38 44 41 41 67 47 61 67 62 54 73 65
DPb
Id A -___
LinkageC
Backbone
Backbone
Sulfate (mol/mol of GlcN) N-
0-
11 11 5.2 10.5 9 11 11 17
16 16 10 16 17 16 18 31
6 6 10 6 6 6 5 3
19 10 17 21 36 28 35 50
75 84 73 73 58 66 60 41
0.98 0.96 0.86 0.89 0.81 0.86 0.74 0.83
1 1 1 1 1 1 1 0
48 14 12 41 04 56 06 86
6.2 16 14 15.5 17 16 27
12 33 28 31 32 32 58
9 3 3 3 3 3 2
39 64 70 56 57 60 63
52 33 27 41 40 37 35
0 63 0.34 0.48 0 49 0 56 0 51 0 32
0 0 0 0 0 0 0
64 38 42 53 82 48 22
96 13
4 81
0.06 (1 . O O ) d
73 7
3-4
6-8 1
Anticoag Act. (IU/mg> 180 109 177 114 152 176 132 16 7
